2018 Pentathlon Race thread

[ao_ika_red]

Erm guys, I have a question. While running NumberFields, my GPU is supposed to be idle, but somehow its temp is 3-4°C higher than the usual. GPU-Z also shows constant spike on GPU load and memory controller load (all those happen in idle clock). Can somebody explain this occurence?

 

[zzuupp]

Erm guys, I have a question. While running NumberFields, my GPU is supposed to be idle, but somehow its temp is 3-4°C higher than the usual. GPU-Z also shows constant spike on GPU load and memory controller load (all those happen in idle clock). Can somebody explain this occurence?

What GPU projects are your crunching when you see this??

 

[ao_ika_red]

What GPU projects are your crunching when you see this??

Just NumberFields, nothing else.

 

[StefanR5R]

While running NumberFields, my GPU is supposed to be idle, but somehow its temp is 3-4°C higher than the usual.

Maybe the GPU fans are pulling air which already went through the CPU cooler.

GPU-Z also shows constant spike on GPU load and memory controller load (all those happen in idle clock). Can somebody explain this occurence?

On Nvidia GPUs, the program

C:\Program Files\NVIDIA Corporation\NVSMI\nvidia-smi​

shows which programs are currently accessing the GPU. I don't know whether an equivalent tool exists for AMD GPUs.

 

[ao_ika_red]

Maybe the GPU fans are pulling air which already went through the CPU cooler.

It goes passive every time it's idle, so no fan involved here. I even paused NumberFields and tried running some leftover Seti CPU tasks, and slowly GPU idle temp returned to normal.

 

[StefanR5R]

The only GPU host of mine, among the hosts which are running NumberFields now, is looking like this:

Spoiler: nvidia-smi output

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 384.111                Driver Version: 384.111                   |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX 108...  Off  | 00000000:01:00.0 Off |                  N/A |
| 23%   31C    P0    57W / 250W |      0MiB / 11158MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX 108...  Off  | 00000000:02:00.0 Off |                  N/A |
| 23%   30C    P0    58W / 250W |      0MiB / 11172MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX 108...  Off  | 00000000:03:00.0 Off |                  N/A |
| 23%   29C    P0    57W / 250W |      0MiB / 11172MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
                                                                           
+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

(Hmm, that's an unholy idle power consumption. As can be seen in the last line, I even shut X11 down.)

(Edit: It's quantum mechanics. Maybe the GPUs were in P12 performance state, maybe not. I read that the act of observing the GPU with nvidia-smi brings it up into P0 state.)

 

[ao_ika_red]

(Hmm, that's an unholy idle power consumption. As can be seen in the last line, I even shut X11 down.)

To put into perspective how unholy it is, your idle power will run my undervolted GPU in full speed.

Maybe the GPU fans are pulling air which already went through the CPU cooler.

I rethink about this and realise that I replaced my CPU fan recently from Noctua Redux S12B to Noctua iPPC F12-3000rpm. In my theory, this new fan draws more air directly from intake fan into the HSF and exhaust case fan thus lesser air flow through the passively cooled GPU. And because my GPU uses conventional aluminium heatsink instead of heatpipe based, it becomes more sensitive to case air flow.

 

[StefanR5R]

2nd discipline:

Swimming: Universe@Home
from 05 May 2018 00:00 UTC
to 12 May 2018 00:00 UTC​

I updated post #3.

 

[Ken g6]

So, how stable is the Universe lately? I heard it might not be very stable.

 

[StefanR5R]

According to your link, the odds are that it will make it through this contest at least.

 

[TennesseeTony]

Universe! Yeah! I already have 1/2 a million points waiting for validation! (quit running it 7? days ago...)

 

[StefanR5R]

As far as I have seen, the Universe validator had several periods lately in which it sat there and did nothing for days. I hope this was only to make it well rested for the Pentathlon.

 

[Pokey]

Am I correct in my observation that both these projects are CPU only?
Next, since I am light on CPUs, and won't be bunkering, I probably should just focus on one of these projects. Any one have a strategic suggestion?

 

[StefanR5R]

Of the 5 races, the only GPU race will be "Cross Country" (5 days), date and project not yet known.

OK, there is a small chance that Asteroids@home will be chosen as "City Run" (5 days), which would be a project in which GPUs could be used too if performance/Watt is of lesser concern...

 

[Skillz]

Universe@home can use ATI GPUs right?

 

[StefanR5R]

They have no tasks available for their opencl_ati application.
Didn't have any for a very long time now.

Edit: typo

 

[Skillz]

thanks for the info

 

[TennesseeTony]

It sure is quiet around here. The BIG EVENT starts in just 27 hours!

 

[Howdy]

It sure is quiet around here. The BIG EVENT starts in just 27 hours!

shhhhhhhhhhh you will disturb my bunkers..........might cause a premature uploadisation!!!

 

[TennesseeTony]

I mean, it's so quiet, it's almost as if we had some sort of secret, private channel to plan out our attack or something.....

IF that is the case, we still need to present a public face, ya know?

 

[ao_ika_red]

I mean, it's so quiet, it's almost as if we had some sort of secret, private channel to plan out our attack or something.....

IF that is the case, we still need to present a public face, ya know?

It's a "calm before the storm", or something, idk.
I just put this for now.

 

[biodoc]

Bunkering, bunkering and more bunkering.

Just under 14 hours to go.

 

[ao_ika_red]

https://www.seti-germany.de/boinc_pentathlon/

from 05 May 2018 00:00 UTC
to 19 May 2018 00:00 UTC​

Rules: It's a competition between teams (not between individuals). There will be 4 concurrent CPU races (one 14 days long, another one 7 days long, a 5er, and a 3 day sprint) and 1 GPU race (5 days). The projects will be made known no earlier than 3 days in advance (for three of the races, 5 days in advance for two races), at random times of the day, together with the start dates.

Points will be given to each team according to their score in each of the 5 races; the final ranking is by the sum of the points. SETI.Germany's Pentathlon site will provide extensive stats with hourly updates.

Outcomes of the Pentathlon in previous years are rather buried at the SETI.Germany site now.

The first discipline was announced on 30 April 00:00 UTC:

Marathon: NumberFields@home
from 05 May 2018 00:00 UTC
to 19 May 2018 00:00 UTC
— stats —​

The 2nd discipline was announced on 2 May 18:00 UTC:

Swimming: Universe@Home
from 05 May 2018 00:00 UTC
to 12 May 2018 00:00 UTC
— stats —​

Note on UTC conversion:

10:00 AEST, 07:00 WIB, 02:00 CEST, 01:00 BST, 00:00 UTC on day X,
20:00 EDT, 19:00 CDT, 18:00 MDT, 17:00 PDT, 14:00 HST on day X-1.​

Note on start time:

A race starts only after the Pentathlon site was able to pull initial stats from the project server. Do not upload your bunker as long as the stats page says "Waiting for initial update. Credits that are granted now will not be counted". Wait until the stats page is showing a table with all teams.​

Edit April 30: marathon announcement
Edit May 2: notes on time conversion and start time

A bit of reminder to everyone here. @StefanR5R covers it really well.

 

[StefanR5R]

Bunkering, bunkering and more bunkering.

Just under 14 hours to go.

That's in the middle of the night for me. So: Even more bunkering, while I get my sleep.

 

[petrusbroder]

52 cores have been bunkering ...
Just 5 hours and 45 minutes until start of the pentathlon ...
I am going to be awake until then ... 01:00 in my timezone ...